Synopsis The Board determined that the crew flew a landing pattern in a narrow valley under marginal weather conditions that dictated an unstabilized approach to the slippery and contaminated runway. The result was a longer-than-normal landing and landing roll. Contributing to the occurrence were a tailwind component, inaccurate James Brake Index (JBI) tables, crew decision making, and the company's use of landing performance tables different from those produced by the aircraft manufacturer for contaminated runways. 1.0 Factual Information 1.1 History of the Flight The British Aerospace BAe 146-200 aircraft, operated by Air BC as flight ABL597, was on a scheduled instrument flight rules (IFR)(1) flight from Vancouver, British Columbia, to Terrace, British Columbia. On board were two pilots, two cabin attendants, and 12 passengers. Prior to the approach into Terrace, the pilots briefed for the localizer/non-directional beacon (LOC/NDB C) approach to Terrace, circling for landing on runway 33 or runway 15. They flew the approach and broke out of cloud about 500 feet above ground level(2) (agl) to the northeast of the airport. They then decided to fly the left-hand circling approach for runway 15 as they did not have the required visual reference to land on runway 33. Due to the poor weather conditions, the crew had to manoeuvre the aircraft around low cloud during short final, but they followed the visual signal provided by the precision approach path indicator (PAPI) lights and established the aircraft on slope before landing. The aircraft touched down on runway 15 about 1,650 feet from the threshold, as measured by airport personnel; it continued down the runway and went off the end at about 25 knots. The aircraft rolled into the overrun area, and came to rest approximately 315 feet from the end of the paved surface, with its nose 39 feet to the right of the extended centre line. The heading of the aircraft at rest was 154 degrees magnetic, 7 degrees right of runway orientation. There was no damage to the aircraft and none of the passengers or crew was injured. The passengers and crew deplaned the aircraft through the forward airstair door and were driven to the airport terminal by bus within about half an hour of the occurrence. The incident took place at latitude 5427'59N and longitude 12834'39W at approximately 0835Pacific standard time (PST)(3) during the hours of daylight. 1.2 Injuries to Persons 1.3 Damage to Aircraft 1.4 Other Damage 1.5 Personnel Information The flight crew were certified and qualified for the flight in accordance with existing regulations. 1.6 Aircraft Information The aircraft was certified, equipped, and maintained in accordance with existing regulations and approved procedures. The weight and centre of gravity were within the prescribed limits. The aircraft was equipped with anti-skid brakes. The aircraft was not equipped with thrust reversers, nor was it required to be by certification standards. The aircraft's main landing gear track is 15 feet 6 inches wide. The aircraft was examined after the occurrence, and no damage, unserviceability, or faults were found that might have contributed to the occurrence. The brakes functioned normally; the main gear tires were examined and there was no evidence of scalloping or scorching such as would be caused by hydroplaning. 1.7 Meteorological Information The weather at the Terrace airport at 0800, 35 minutes before the occurrence, was overcast ceiling measured 500 feet agl, visibility two and one-half miles in light freezing drizzle, light snow and very light ice pellets, temperature minus two degrees Celsius, dew point minus three degrees Celsius, wind from 010 degrees True at seven knots, altimeter setting 29.94inches of mercury. Just before landing, the FSS reported to the crew that the surface wind was from 350 degrees magnetic at five knots. The weather at 0842, seven minutes after the occurrence, was broken ceiling measured 500feet agl, overcast 1,200 feet agl, visibility three miles in very light freezing drizzle and light ice pellets, wind from 020 degrees True at six knots. 1.8 Crew Estimation of Landing Field Length Before landing, the crew referred to charts in the Air BC BAe 146 Quick Reference Handbook to estimate the required landing field length. They used the Landing Field Length - Dry Paved Runway chart to calculate the landing field length for a dry runway, taking the existing tailwind into consideration. The resultant distance was then entered into the Landing Distance Required Corrected For JBI 60% Factor chart. This chart revealed that a runway length of 4,750 feet was required to land with a James Brake Index (JBI) of 0.25, and that the aircraft could land on the Terrace runway with a JBI reading as low as 0.20. These figures assumed the touchdown point to be a maximum of 1,000 feet beyond the runway threshold. Based on this information, the pilots were confident that the aircraft could land and stop safely on the 6,000-foot Terrace runway. 1.9 Approach and Landing The Air BC BAe 146 Pilot's Handbook defines a Stabilized Approach as stabilized airspeed, stabilized sink rate, and a constant profile. The BAe 146 Operations Manual indicates that it is essential to arrive over the threshold at VREF 33 with the thrust levers coming back to flight idle, aiming to touch down about 1,000 feet into the runway at about VREF 33 - 7 knots. The aircraft landed about 1,650 feet beyond the runway threshold, which is about 650 feet beyond the ideal touchdown point. The co-pilot assessed that a stabilized approach to runway 15 at Terrace is difficult, due to the confined space and the high terrain, especially in poor weather. The pilots intended to land into wind, on runway 33, which would have required a small left turn from the approach track, if the runway environment had been acquired prior to the missed approach point (MAP). They had also briefed for a circling approach to runway 15 in the event they were unable to land on runway 33. When the aircraft arrived at the missed approach point, the weather conditions did not permit a landing on runway 33, so the pilots continued with the left circling pattern to runway 15. The co-pilot, who was flying the aircraft from the right-hand pilot's seat, lost sight of the runway during the circling procedure. The captain did not take control of the aircraft; however, he provided the co-pilot with directions to the final approach path. Flight data recorder (FDR) information reveals that the aircraft passed through the runway centre line during the turn from base leg to final approach, and that bank angles up to 24 degrees were required to manoeuvre the aircraft around the low cloud on final approach. The pilots reported that the aircraft was established on the PAPI glide slope before landing. The airbrake was deployed before touchdown, which is normal procedure. The aircraft touched down on the centre line of the runway at about 108 knots, at a point measured 1,650feet beyond the runway threshold. The spoilers were deployed at touchdown. The co-pilot applied the brakes immediately after touchdown; however, they were not effective. The captain realized that the aircraft was not decelerating as expected so he shut down the two outboard engines. The aircraft ran off the end of the runway at about 25 knots. There was about 6 inches of soft snow on the runway overrun, and the aircraft occupants reported experiencing only a slight bump as the aircraft left the runway. Some passengers reported hearing a sound during the landing roll that was identified as the anti-lock brake system cycling on and off. 1.10 Crew Decisions When the crew decided to continue for a left-hand circling procedure to runway 15, the co-pilot continued to fly the approach. However, because the runway was to his left and below his cross-cockpit line of sight, he could not judge the turn required to align the aircraft with the final approach path and was forced to rely on the captain's directions. It is not unusual for the pilot with a clear view of the runway during a circling procedure to take control and land the aircraft. Additionally, during the last stages of the approach, the captain, on three occasions, expressed his doubts about the landing, and suggested that they go around, but he did not insist. The co-pilot expressed doubts about being able to see the runway again if they attempted another approach and continued with the approach and landing. 1.11 James Brake Index and Runway Condition Information JBI is a measure of braking action on a runway surface, with 0.8 being the best value, and 0.0 being no braking action. The JBI was measured with the prescribed Transport Canada (TC) decelerometer in an approved truck at intervals of approximately 1,000 feet, and approximately 30 feet either side of the runway centre line, all in accordance with standard procedures. The equipment used to measure the JBI was examined and tested, and no fault was found. The method of measuring the JBI, and the qualifications of the workers who made the measurements were examined and were found to be in order. Urea is a chemical applied to the runway surface to melt ice. It is applied from the centre of the runway outwards. On the day of the occurrence, the runway clearing operation began shortly after 0600, with the airport crews sweeping the snow off the runway and applying urea. At 0620, runway 15was reported to be 100 per cent ice-covered, with a JBI of 0.26. The pilot of an MU-2, an aircraft with a wheelbase of slightly less than eight feet, landed on runway 33 at 0820, 15 minutes before the occurrence. He reported that the braking action was fair. He said that the runway surface was not slushy but wet with granularity, like wet pavement, and that the runway was wetter and clearer in the middle. At 0828, seven minutes before the occurrence, the runway was reported as 100 per cent ice-covered, with a trace of slush in the centre 60 feet. The average JBI was 0.36 with readings of 0.24/0.29 at the south end of the runway. This information was relayed to the BAe 146 crew at 0829 by the Terrace Flight Service Station (FSS) as runway 100 per cent ice-covered, centre 60 feet trace of slush, JBI average 0.36 down to about 0.29 at the south end of the runway. The value of 0.24 was not included in the report to the flight. At 0842, seven minutes after the occurrence, the average JBI was 0.35, with readings of 0.24at the south end. The BAe 146 captain walked on the south end of runway 15 shortly after the occurrence, and noted that the runway surface at the centre was ice covered with water on it about 1/16 of an inch deep, and that the runway was very slippery. 1.12 Emergency Rescue Services (ERS) Response The ERS crew responded immediately and reached the aircraft within a few minutes. As there was no fire, damage or injury, their services were not required, beyond facilitating the transfer of the passengers to the main terminal. 1.13 Aids to Navigation All aids to navigation were operating normally. 1.14 Terrace Airport Information Runway 15 at Terrace airport is asphalt, has a downslope of 0.28 per cent and is oriented 147degrees magnetic. It is 6,000 feet long and 150 feet wide. The airport is located at the end of a valley and is surrounded on three sides by mountains rising to over 10,000 feet above sea level (asl), and the airport reference point elevation is 713 feet asl. 1.15 Air BC BAe 146 Performance Charts The Air BC BAe 146 Pilot's Handbook, Appendix C-3, states that, when the runway is forecast or reported to be wet or slippery at the arrival airport, then the Factored Wet Landing Distance method applies (Dry Factored Distance plus 15 per cent), unless the estimated or actual JBI is less than 0.28. When the estimated or actual JBI is less than approximately 0.28, the Dry Factored Distance with a JBI correction should apply. This distance can never be less than the wet distance calculated above. The Dry and Wet Runway Landing Field Length charts in the Air BC BAe 146 Pilot's Handbook and the Quick Reference Handbook used by the flight crew were prepared by BritishAerospace, the manufacturer, and accepted by TC during the process of certifying the aircraft to operate in Canada in 1984. At that time, the regulations did not require consideration of landing field length on contaminated runways. The standard procedure in Canada for estimating landing field length on contaminated runways was to apply the factored Dry Runway figures to the JBI tables. The JBItable increased the Dry Runway figures to account for reduced braking efficiency on slippery runways. The Dry Runway figures contain the 60 per cent factor, meaning that the aircraft's demonstrated landing distance was 60 per cent of the stated Landing Field Length distance, as was required by regulation. For example, if the landing distance for a Dry Runway Field Length is 5,000 feet, then the aircraft can stop in 60 per cent of this distance, or 3,000 feet. The remaining 2,000 feet is a safety margin to account for differences in pilot technique, and for less than ideal runway conditions. When the Dry Runway Field Length is modified in the JBI tables for reduced braking coefficients, the 40 per cent safety margin is already used up, hence the JBI table does not start until the JBI is 0.35 or less, that is, when the aircraft has already passed into the 40percent safety margin. The Landing Runway Field Length figures in the JBI tables, therefore, are minimum stopping distances with no safety margin. Stopping within these distances requires a textbook landing, that is, an ideal approach path and touchdown, correct approach speed, and prompt application of maximum braking and lift dumping devices. Any piloting or performance factors that are less than ideal will result in a landing distance longer than that which appears in the JBI table. The 60 per cent factor is not explained in the Aeronautical Information Publication Canada (AIP). During their training, Air BC BAe 146 pilots are not taught that the 60 per cent factor is not present in the JBI tables, and their training leads them to use the JBI figures as accurate predictions of aircraft landing performance on contaminated runways. Air BC reports that, in its six years of experience operating the BAe 146 on contaminated runways, the aircraft has performed according to the landing and JBI charts. The company believes that the aircraft went off the end of the runway because the JBI at the centre of the runway was less than where it was measured 30 feet either side of the centre line. 1.16 JBI Tables TSB Engineering Branch report LP 44/94, Calculation of Stopping Distances on Contaminated Runways using James Brake Index (JBI), concluded that: The method used to measure JBI is reasonable. The JBI table in the AIP Canada should contain a warning that the distances assume a textbook landing, and do not contain the safety margin of the 60percent factor. The calculation of the JBI figures does not consider the reduction in braking efficiency at lower JBI values, and is non-conservative. The JBI table was created in 1973; however, TC has no engineering data on record to validate the JBI figures. A 1993 TC study entitled The Validity and Applicability of the James Brake Index to Aircraft Landing Performance was conducted to substantiate the JBI figures. The report concluded that actual landing distances may be as much as 18 per cent greater than the JBI table figures, and that the 60 per cent safety factor is removed. At the time of the occurrence, no amendment had been made to the JBI information in the AIP. The AIP, section AIR 1.6.4, states that JBIs are accurate only for packed snow or ice, and not for wet or slush-covered runways. It also notes that, while there has been insufficient flight test verification to warrant formal approval, the information should prove a useful guide to pilots in estimating aircraft performance under adverse runway conditions. The explanation below Table B, Landing Distance Required Corrected for JBI and 60%Factor, in section AIR 1.6.6 of the AIP states as follows: ...it is designed for aircraft where performance data contains the 60% factor required by regulation. The table does not contain information for JBI numbers above .4 because the 60% factor already applied to the flight manual data will cater to runway requirements for JBI numbers above .4. Thus, although the safety margin of the 60 per cent factor is not present in the JBI table, this is not clearly stated in the AIP. There is no explanation or definition of the 60 per cent factor in the AIP. Discussions with users of the JBI tables indicate that not everyone was aware that the 60percent factor is not present. The Air BC BAe 146 Pilot's Handbook warns pilots that charts and graphs are only part of the solution when calculating landing distances, and states that pilots should remain situationally aware, which means they should consider other factors that influence landing on a wet or slippery runway. They are also warned of the importance of flying the approach on profile at the correct speed, so as not to consume valuable runway. It is also stated that all landing distances derived from the JBI tables are considered estimated distances and are only intended as a guide. There is no direction on how to use these estimated distances as a guide when calculating landing distances. There is no mention in this handbook that JBIfigures are only valid for packed snow and ice, as stated in the AIP, or that the safety margin of the 60 per cent factor is gone. Certification is not predicated on the use of reverse thrust when calculating landing distances; but, in fact, most large category aircraft are equipped with thrust reversers, and pilots of these aircraft use them when landing. The use of reverse thrust during the landing phase represents an extra stopping performance margin on slippery runways which is not available to the BAe 146 aircraft. 1.17 Information from the Manufacturer British Aerospace (BAe), the manufacturer, was consulted regarding the occurrence. They stated that the JBI method of evaluating the BAe 146 performance on a slippery runway is not acceptable, and is totally inconsistent with the BAe figures. They refer to the BAe 146 Manufacturer's Operations Manual (MOM), section 9.10.11, Handling Normal, Ice and Rain, which reveals that, given the circumstances on the day of the runway excursion, the aircraft would have required a ground roll of about 4,550 feet. When this distance is added to the 1,650 feet of runway overflown by the landing aircraft, a runway about 200 feet longer than the 6,000-foot Terrace runway was required. Furthermore, the landing distance required would have been increased by the tailwind component and the downslope of the runway. The information in MOM section 9.10.11 is repeated in the BAe 146 Aircraft Flight Manual and was available to TC during the BAe 146 certification process; it was made available to Air BC in the fall of 1993. There is a disparity between the landing performance figures of Air BC and those of BAe. 1.18 Flight Recorders There were two flight recorders installed on this aircraft at the time of the incident: the flight data recorder (FDR) and the cockpit voice recorder (CVR). Both units were removed from the aircraft, and the TSB retrieved and analyzed the data. The FDR revealed that the aircraft was not in a stabilized approach before landing. The aircraft passed through the extended runway centre line as it turned onto final approach because the crew had to manoeuvre around low cloud, and angles of bank of up to 24degrees were recorded when it was on short final. FDR data indicated that the aircraft touched down, left main gear first, 1,500feet from the runway threshold(4), and the spoilers deployed at the 1,900-foot point. The captain shut down the outboard engines 4,200 feet from the threshold. The aircraft ran off the end of the runway at a speed of approximately 25 knots, about 42 seconds after touchdown. The target final approach speed was 118 knots; the FDR indicated that the actual final approach speed was between 114 and 120 knots. The target flare speed was 113 knots; the FDR indicated that the actual flare speed was 117 knots. The target touchdown speed was 106 knots, and the FDR indicated that the actual touchdown speed was 108 knots. The CVR contains an endless loop of tape which records, in this case, 33 minutes and 39seconds of the most recent cockpit communications. Since the CVR had remained powered until the crew shut down the aircraft electrical power about 32 minutes after landing, only the last 1 minute and 39 seconds of this flight's communications remained for investigation purposes.